Expanding the Boundaries of
Business Intelligence
The New Era of BI Goes Far Beyond Data and Reporting
Until recently, business intelligence (BI) applications were used primarily
to collect and analyze historical data. Analytical tools, concentrated in
the hands of just a few well-trained users, produced reports for high-level
strategic decision making. Now the role of BI is expanding as businesses
are looking to collect more information, analyze it quickly, distribute it to
multiple users and deliver it back into transactional applications to tune
core business processes in real-time.
To support this expanding vision, many businesses are working to
consolidate their fragmented BI systems and applications. This integration
effort allows the BI infrastructure to support more users while reducing
costs. Dual-Core Intel® Itanium® processors running Windows or Linux
offer a much needed resource for this consolidation. They provide the
high-end scalability and availability of traditional RISC and mainframe
architectures, as well as the flexibility of standards-based servers and
widely supported operating systems that help organizations reduce their
total costs.
Moving Toward Ubiquitous Business Intelligence
“The new era of BI, which is already here, goes far beyond data and
reporting. BI is becoming proactive, real-time, operational, integrated
with business processes and extending beyond the boundaries of the
organization,” says Neil Raden, Intelligent Enterprise.¹
Most business and IT decision makers are well aware of the enormous
value locked inside enterprise data and are focused on tapping that
potential for competitive advantage. Yet for most large businesses, moving
forward means first taking a step back to assess existing solutions and
new opportunities.
Anatomy of a Business Intelligence System
BI solutions are complex and typically include a number of components
running on different systems (Figure 1). These systems contain valuable
information which IT organizations identify, collect, reformat and cleanse.
Figure 1. A Large BI Implementation
¹ Source: Business Intelligence 2.0: Simpler, More Accessible, Inevitable, by Neil Raden, Intelligent Enterprise, February 2007.
http://www.intelligententerprise.com/channels/bi/showArticle.jhtml?articleID=197002610.
Then IT loads the data into a staging database, a data warehouse and
possibly multiple data marts for analysis. This process, known as ETL
(Extract Transform and Load), helps ensure the data is validated and
integrated, enabling efficient analysis and a single-version-of-the-truth.
The data warehouse is the foundation of BI, and the point where server
workloads are most demanding. In today’s business environment, data
warehouse and ETL processes need to support large data volumes (which
typically double every year), global environments and 24x7 operations.
Not only must businesses integrate more data, but in many cases, they
must load it faster to deliver more timely information to decision makers.
As a result, batch update windows are shrinking. Many IT organizations are
already finding they no longer have the luxury of nightly or weekly updates,
but must load new data into their warehouses within a few hours.
Re-Engineering Business Intelligence Systems
Today, BI capabilities are expanding. Businesses are evaluating and
deploying various strategies that can increase the value of BI systems
and decrease infrastructure cost, including:
• Consolidate BI systems
• Increase software reuse and integration
• Widely distribute information and analytical tools
• Improve feedback mechanisms
A Business Intelligence Platform for a New Era
The transition to a more versatile and expanding business intelligence
system requires a high performance computing platform that can
deliver extreme performance, scalability, availability and flexibility. The
underlying server infrastructure is well served by today’s Dual-Core Intel
Itanium processors, which deliver mainframe-class capabilities and meet
crucial system requirements:
Scalability:
• Per Cluster: Up to 10,000+ processors per cluster (Linux)
• Per Server: 512 dual-core processors with 128TB of globally
shared memory
• Per Processor: 1,000 TB memory addressability, 24MB cache and
8GB/s I/O bandwidth
Mainframe-class Availability:
• Advanced error detection and correction across all major data paths
• Industry-leading innovations, such as core-level lockstep
• Mainframe-class systems available from multiple vendors
Flexibility through a Widely Supported, Standards-based Architecture
• Supported by 9 of the top 10 server vendors
• Multi-OS support: Windows, Linux, UNIX and more
• More than 12,000 applications from more than 2,000 ISVs

Figure 2. Itanium-based Systems Offers A More Flexible, Affordable & Widely
Supported Hardware & Software Platform
Industry-Leading Benchmark Results
Itanium-based servers were specifically designed to handle the large
data volumes and complex transactions of BI applications. They support
multiple terabytes of physical memory and high-performance Dual-Core
Intel Itanium processors with up to 24MB of on-die cache, more than any
other current processor architecture. Itanium-based servers are delivering
world-record performance for today’s most demanding, data-intensive
applications (Table 1).
Workload Benchmark
Data Warehouse TPC-H Benchmark @ 30,000GB
OLTP (Online Transaction Processing) TPC-C Benchmark b (tpmC)
ERP (Enterprise Resource Planning) Two-tier SAP SD b Benchmark (Users)
Java Server SPECjbbb 2005 (BOPS)
Table 1. High-End Benchmarks For Dual-Core Itanium Processor-Based Servers
High Availability for Mission-Critical Support
As BI applications stretch beyond their traditional roots to link directly with
transactional applications, they are beginning to have a direct impact on
revenue streams and operational efficiency. With this evolution, the cost
of downtime is increasing. “It is no longer acceptable to have a warehouse
with a lower standard of availability, fault tolerance or disaster recovery
than operational systems,” says Mark A. Beyer, Gartner Analyst.²
Itanium-based systems are well-suited for these mission-critical
applications, delivering the kind of per system stability and high-
availability typically associated with high-end RISC and mainframe
systems. Extensive capabilities for error monitoring, containment and
correction are built into the Dual-Core Intel Itanium processor, along with
an Enhanced Machine Check Architecture that provides a standards-
based foundation for integrated, system-wide error management.
The Road Ahead
BI has been around for more than 30 years. It has proven its strategic
value, yet enormous potential remains untapped. Future innovations will
a Performance tests and ratings are measured using specific computer systems and/or components and reflect the approximate performance of Intel products as measured by those tests. Any difference
in system hardware or software design or configuration may affect actual performance. Buyers should consult other sources of information to evaluate the performance of systems or components they are
considering purchasing. For more information on performance tests and on the performance of Intel products, visit http://www.intel.com/performance/resources/limits.htm
b World Record Performance Source / Vs. Competition: http://www.ideasinternational.com/benchmark/ben010.aspx as of 20 March 07. Actual performance may vary. See website for configuration details.
c The benchmark results shown are world records or #1 for the specified OS as of 20 March 07.
² Source: Key Issues for Data Warehousing, 2007, by Mark A. Beyer, Gartner, March 30, 2007, ID
Number: G00147102.
be directed at making it easier to use, less costly and more pervasive.
“…a vast population exists whose business intelligence requirements
have not been met to their full potential. The next wave of BI will reach
out to these employees as well as other organizational stakeholders such
as suppliers, partners, customers, and government agencies….”
— IDC, Worldwide Business Intelligence Tools 2005 Vendor Shares.
To take advantage of these developments, most businesses will need:
• A strategic, enterprise-wide approach to BI to avoid tactical, stovepipe
implementations that add complexity and limit options.
• A better understanding of enterprise processes, and the points where
BI can be used to improve them.
• Standardization of enterprise master data and metadata as part of
broader data quality and compliance effort.
• Strategies for identifying and integrating “shadow BI” solutions, such
as spreadsheets, that can thwart attempts to achieve a single-version
of-the-truth.
• Better use of SOA to populate data warehouses from a wider variety of
sources and to deliver information where, when and how it is needed.
Result a Status c
150,960.5QphH @ 46.69$/QphH World Record
4,092,779 World Record
30,000 World Record
4,231,610 World Record
• Simpler and more pervasive analytical tools and distribution
mechanisms that give more users access to better information that is
pertinent to their needs.
As business intelligence solutions evolve, many companies are making
system changes that require higher levels of performance, scalability and
availability from the BI infrastructure and greater flexibility for extending
solutions as change continues.
Itanium-based servers running Windows or Linux offer a high-performance
computing solution for these growing requirements. They deliver the
scalability and availability of traditional RISC and mainframe systems,
but on a more flexible, affordable and widely supported hardware and
software platform. As businesses continue to extend their BI solutions,
these advantages will be valuable in helping them keep pace with
technological change and growing business demands.
For more information on Itanium-based solutions for business intelligence,
please visit www.itaniumsolutions.com/apac